Discussion of Monetary Policy Under Uncertainty in Micro-Founded Macroeconometric Models, by Levin, Onatski, Williams and Williams

The marriage of micro-founded dynamic stochastic general equilibrium models and nominal rigidities has opened up new possibilities for policy analysis, and in particular for conducting welfare-based analysis using estimated macro models. Andy Levin, Alexsei Onatski, John Williams, and Noah Williams (henceforth LOWW) have exploited these possibilities in an excellent paper, one in which they accomplish three things. First, they specify a second generation new Keynesian model and estimate it using Bayesian techniques. Second, they use this model to calculate the welfare costs of fluctuations under alternative policies. Third, they investigate the robustness of policy rules to parameter and model mis-specification. Each of these tasks is carried out in elegant fashion, producing a paper that not only represents the state of art in modern monetary economics, but provides an outstanding example of how to do policy analysis. But because LOWW provide such a carefully and well-specified model, one that includes all the bells and whistles of the latest generation of new Keynesian models, it offers an opportunity for assessing the ability of these models to serve as the workhorse for monetary policy analysis. I will argue that this estimated new Keynesian model fits the data in a very non-Keynesian manner. By this I mean the model attributes most of the actual output ∗University of California, Santa Cruz, [email protected].